Different from an existing frequency division duplex (FDD for short) or time division duplex (TDD for short) technology, a wireless full-duplex technology allows receiving and sending operations simultaneously performed on a same wireless channel. Therefore, theoretically, spectral efficiency of the wireless full-duplex technology is twice as much as that of the FDD or TDD technology. A precondition for implementing wireless full-duplex is to avoid, reduce, or cancel self-interference caused by a transmit signal of a transceiver to a receive signal of the transceiver as much as possible, so as not to adversely affect correct reception of a wanted signal.
In an existing wireless full-duplex multiple-input multiple-output (MIMO for short) system, a strength of a self-interference signal obtained after space isolation performed by using a transmit and receive antenna is still far greater than a strength of a wanted signal. Therefore, generally, a data interference cancellation operation needs to be performed on the self-interference signal by combining a radio frequency self-interference cancellation (RF SIC for short) technology and a digital self-interference cancellation (Digital SIC for short) technology. In both a radio frequency interference cancellation process and a digital interference cancellation process, a self-interference channel needs to be estimated, that is, a channel between each transmit antenna and each receive antenna of a same transceiver. Generally, dedicated downlink transmission may be performed by using some half-duplex timeslots. A self-interference signal is received by using each receive antenna, and self-interference channel estimation is performed according to a received downlink signal. Specifically, each transmit antenna may perform downlink transmission by using a non-orthogonal time-frequency resource or an orthogonal time-frequency resource. Correspondingly, each receive antenna receives non-orthogonal self-interference signals or orthogonal self-interference signals from different transmit antennas. For a non-orthogonal self-interference signal, for example, in an orthogonal frequency division multiplexing (OFDM for short) system, if a quantity of transmit antennas is N, each receive antenna needs to solve N linear equations on each subcarrier to obtain a self-interference channel. For an orthogonal self-interference signal, the transmit antennas use orthogonal time-frequency resources. Therefore, each receive antenna does not need to solve equations, and performs channel estimation on each orthogonal time-frequency resource, so as to implement the self-interference channel estimation.
However, in the prior art, for a non-orthogonal self-interference signal, it is highly complex to implement the self-interference channel estimation. Therefore, it is also complex to perform interference cancellation on the non-orthogonal self-interference signal. For an orthogonal self-interference signal, the orthogonal self-interference signal is sent on an orthogonal time-frequency resource. Therefore, a resource is poorly utilized during interference cancellation performed on the orthogonal time-frequency resource.